The construct validity of the Swedish MHLC-C was supported by little, if any, inter-correlations between the internal and external sub-scales, which are intended to measure separate dimensions of health locus of control. The low correlations found between ‘Other people’ and ‘Doctors’ and ‘Chance’, respectively, are probably due to their original inclusion in the same external subscale of MHLC-A (Wallston et al 1978). Our finding of little, if any, correlations between the MHLC-C and most other
constructs supports health locus of control as a unique construct that does not correlate with demographic or disease-related factors. The low correlations found between external health loci of control and learned helplessness were expected as these constructs may, to some extent, target similar beliefs.
The test-retest stability of the proposed Swedish MHLC-C was satisfactory, and generally better than that demonstrated for the original MHLC-C (Wallston et al 1994).
The homogeneity of the items within each sub-scale was good for all but 'Other people', where it was somewhat low.
Non-parametric statistical procedures are often recommended for analysis of ordinal data. However, the measurement properties of the original MHLC-C were analyzed with parametric statistics. Thus, parallel analyses were carried out in the present work to analyze the measurement properties of the Swedish MHLC-C. As the analyses did not reveal any major differences, the outcome of the parametric analyses is presented in order to improve comparisons with the data from the original publication on MHLC-C (Wallston et al 1994).
5.3 HEALTH LOCUS OF CONTROL
The patient’s attribution of health locus of control mainly to ‘Doctors’ (Studies II-IV), or rather health professionals, seems adequate for patients diagnosed with a life-long illness requir ing constant monitoring and treatment in health care. This attribution also accords with a previous study of samples of patients in the USA with long-standing RA and other long-standing illnesses (Wallston et al 1994).
The finding of only weak correlations between higher age and the MHLC-C ‘Doctors’
(Study III) was somewhat unexpected as others have found stronger correlations between higher age and all the MHLC-C subscales (Hashimoto & Fukuhara 2004, Kuwahara et al 2004). However, those studies were not conducted in patients with RA.
Thus, the present findings may indicate that factors related to the disease itself are more powerful than age and gender as to relation with health locus of control.
Individuals have often been expected to possess either internal or external health locus of control beliefs. This was, however, not the intention with the construct (Rotter 1990).
Rather, one individual may possess high or low beliefs in both internal and external control over their health, which may be particularly applicable to those with long-standing illnesses that require regular treatment and monitoring in health care. In a study of 159 Chinese women, 25 were pure ‘internals’, 36 pure ‘externals, 37 ‘duals’, and 61t ‘no-control believers’ (Wu et al 2004). In the present work (Study IV) only one patient was purely ‘internal’ and one purely ‘external’, while 93 where either high or
scarce use of the MHLC-C within health care. Only a few older studies on health locus of control in patients with RA have been published, which might indicate that this construct has not been considered applicable in this context.
5.3.1 Pain and health locus of control
A relation between high internal health locus of control and high intensity pain in patients with early RA was found in the present work (Study III), while no relations between MHLC-C subscales and pain were found among those with long-standing RA (Study IV). One reason might be that patients with early RA stay more active and therefore tolerate higher pain because the positive reinforcements of physical activity rule out the potentially negative reinforcement of pain. This hypothesis is in line with the health locus of control construct (Rotter 1975).
5.3.2 Physical activity and health locus of control
As a relation was found between moderate/high physical activity and high attribution of health locus of control to health professionals among patients with early RA, it was unexpected that no such relation existed among patients with longstanding RA. Thus, the result indicates that such relations may depend on disease duration. A possible explanation for this might be the trust that recently-diagnosed patients may have in the specially-trained health professionals at rheumatology clinics and their
recommendations to be physically active. Another explanation might be that patients with shorter disease duration have less disability, which makes it easier to follow such recommendations.
Others have found poorer health behavior among individuals with high external health locus of control (Smith & Mason 2001, Conant 1998, May et al 1998), while betterhealth behavior has been found among those with high internal control (Unger 1997). However, none of these studies addressed patients with RA. One reason for the absence of
relations between physical activity and internal health locus of control in the present work might be that, particularly among those with early RA, in their new roles as people with disease, they lack the confidence to decide whether physical activity is beneficial for their arthritis. This presumption also fits with the finding of a strong relation between adequate physical activity and high attribution of health locus of control to health professionals.
5.3.3 PAIRS and fear avoidance beliefs
The strong relation between high-intensity pain and strong beliefs in a relationship between pain and impairment found in the present work (Study IV) may not seem surprising. It rather seems natural that those with much pain hold such beliefs, as they may experience constant consequences of their pain in daily life. However, another explanation might be that individuals holding such beliefs are more likely to develop or experience more pain. This needs to be investigated in future prospective studies.
The strong relation between high pain intensity and strong fear-avoidance beliefs (Study IV) is in accordance with previous studies investigating fear-avoidance beliefs as predictors of chronic low-back pain in patients with lumbar and cervical spine pain (Picavet et al 2002, George et al 2001).
5.3.4 Limitations of the present work
More emphasis on strategies to maintain new skills and use them in daily life might have improved the results of the relaxation training program. The inclusion of patients with more disability, preferably related to muscle pain, might also have been adequate as many of the present patients had very little disability and thus not much room for improvement following the training program. Further, personal instructions during the training sessions and some kind of encouragement during the follow-up period would have improved results. Finally, the fact that the study was not conclusive and that mood and tension were not assessed means that the relaxation training program might have conferred further improvements than those demonstrated.
The choice to investigate test-retest stability of the MHLC-C with two sets of data, one obtained in the waiting room and one at home, might have influenced the results.
However, as health beliefs could be expected to show some trait-like stability (Wallston et al 1994), this might have been a minor problem. Opinions about the size of
correlation coefficients indicating meaningful relationships seem to differ between traditions in medical and behavioral-science literature. In Study II, a more behaviorally oriented way of interpretation was chosen, which may of course be questioned.
As to the external validity of the present work, the participating patients were recruited somewhat differently for the different studies. While there is no reason to suspect systematic bias in the selection of the participating clinics or patients, all samples may be considered as such of convenience. Hence, those who accepted to participate may have had more positive health beliefs and a greater interest in their health than non-participants had. While such problems cannot be avoided as long as study participation is optional, it is still important to bear in mind that the results of the present work can only be generalized to those patients with RA who have similar characteristics.
5.3.5 Clinical implications
Muscle relaxation training with taped instructions improves some aspects of muscle function and self-rated health. However, results might be better with supervised training and the application of strategies for implementing the acquired skills in daily life.
The MHLC-C might be useful in physical therapy for surveying health cognitions among patients with early RA in order to improve strategies for pain reduction and healthy exercise behavior. The PAIRS and the mFABQ might be useful in physical therapy to survey health cognitions among patients with RA in order to improve strategies for pain reduction and healthy exercise behaviors.
5.3.6 Future research
Future studies of relaxation training programs should include evaluation of sleep and contain elements of coaching and follow-up sessions in order to motivate patients to implement the acquired skills in daily life. Another issue of interest to investigate is whether introduction of relaxation training at an early stage of RA would prevent or limit the storage of memories of pain deep in the thalamus, and thus minimize the risk of vicious circles of pain.
Further investigation of the responsiveness and sensitivity to change of the MHLC-C is needed. Beliefs about relationships between pain and impairment and beliefs about fear-avoidance of activity need further investigation in order to develop clinical
strategies for reduc ing such beliefs in patients with RA. Education and graded exposure to movement have been suggested as one way forward in work with patients with long-standing pain who report high fear-avoidance beliefs (Vlaeyen & Linton 2000, Burton et al 1999). Whether this also applies to patients with RA needs to be studied. Further studies are also needed to explore cognition in relation to physical activity among patients with RA.